ORIGINAL ARTICLE

Anal Skin Tags in Inflammatory Bowel Disease: New Observations and a Clinical Review

Jennifer L. Bonheur, MD, Jared Braunstein, DO, Burton I. Korelitz, MD, and Georgia Panagopoulos, PhD

observations support the diagnostic significance of AST heralding Background: The association between intestinal Crohn’s disease the diagnosis of CD when they are discovered on physical exam, (CD) and specific perianal abnormalities called anal skin tags (AST) especially in young people with , abdominal pain, and/or has been recognized but not well defined. Skin tags have been growth retardation. classified into 2 types: 1) raised, broad, or narrow, single or multiple, soft or firm, and painless, often referred to as “elephant ears”; or 2) (Inflamm Bowel Dis 2008;14:1236–1239) edematous, hard, often cyanotic, tender or not, and typically arising Key Words: anal skin tags, elephant ears, IBD, Crohn’s disease, from a healed anal fissure, ulcer, or . The aims of this ulcerative , other perirectal manifestations study were to i) better characterize those skin tags identified by the term “elephant ears” and differentiate them from other types of AST; ii) compare their prevalence in patients with CD and ulcerative he association between intestinal Crohn’s disease (CD) colitis (UC); iii) observe the relationship of the skin tags to the Tand characteristic anal skin tags (ASTs) has been recog- location of the primary bowel disease; and iv) to discuss the value of nized but not well defined. Characteristic morphologic fea- these typical AST in making an early diagnosis of CD. tures of AST were studied. An attempt was made to divide Methods: Photographs of all AST were taken when present at them into 2 types (Table 1; Fig. 1A,B): 1) those often referred lower endoscopy in 170 consecutive patients with inflammatory to as “elephant ears,” and 2) those typically arising from a bowel disease (IBD) seen in the private office of the senior inves- healed anal fissure, ulcer, or hemorrhoid.1 Type 1 are more tigator and Lenox Hill Hospital. Data was gathered with respect to typically characteristic of CD in contrast to major differences between the 2 types of AST. The location of the (UC), although type 2 may also be seen in either disease or primary bowel disease for these patients was obtained from an may overlap in appearance with typical painless external extensive IBD computer database and review of details from charts. , prolapsed internal hemorrhoids, and occasion- Results: Specific features of AST were described and served to ally with type 1 whether occurring in association with CD or 1,2 favor type 1 versus type 2. AST were found more frequently in not. patients with CD (75.4%) as compared to patients with UC (24.6%), Our purpose was 1) to better characterize those AST confirming previous observations that they are more diagnostic of identified as “elephant ears” typically associated with CD; 2) CD (P ϭ 0.005). Subset analysis revealed a trend with a greater to differentiate them from other AST more likely to have incidence of AST in patients with colitis (46.9%) as compared to arisen from fissures, ulcers, or hemorrhoids; 3) to compare patients with (36.7%) and ileocolitis (16.3%) (P ϭ 0.067). their prevalence in patients with CD and UC; 4) to compare location of the primary bowel disease in patients with AST; Conclusions: We provide photographs with the most characteristic and 5) to discuss the value of these typical AST in making the features of AST and attempt to separate elephant ears (type 1) from less typical AST (type 2) in CD. Our study confirms previous reports early diagnosis of CD. that AST are more commonly found in association with CD as compared with UC and more so in the presence of disease limited to MATERIALS AND METHODS the colon as compared to disease elsewhere in the bowel. Our This study had the approval of the Institutional Review Board (IRB) at Lenox Hill Hospital. AST were photographed in 170 consecutive patients with inflammatory bowel disease (IBD) Received for publication November 14, 2007; Accepted February 21, and AST in the course of 1 year and details were collected 2008. From the Department of , Lenox Hill Hospital, and New from our database and their charts to determine diagnosis York University School of Medicine, New York, New York. (CD versus UC) based on clinical, endoscopic, radiological, Reprints: Jennifer L. Bonheur, MD, Department of Gastroenterology, and histological features independent of the presence of skin Lenox Hill Hospital, 100 E. 77th St., New York, NY 10021 (e-mail: tags. Photographs of anorectal protrusions suggesting AST [email protected]). were taken at colonoscopy or flexible sigmoidoscopy by the Copyright © 2008 Crohn’s & Colitis Foundation of America, Inc. DOI 10.1002/ibd.20458 senior investigator (B.I.K.) and then were gathered and sub- Published online 1 May 2008 in Wiley InterScience (www.interscience. sequently reviewed as a group to better determine specific wiley.com). morphologic characteristics. The differential diagnosis of

1236 Inflamm Bowel Dis ● Volume 14, Number 9, September 2008 Inflamm Bowel Dis ● Volume 14, Number 9, September 2008 Anal Skin Tags in IBD

TABLE 1. Characteristic Features of Anal Skin Tags of Crohn’s AST made throughout this study was initially made by the Disease senior investigator. Close attention was paid to better char- acterize the morphologic characteristics of AST associated Type 1 Type 2 with CD (“elephant ears”) as compared to those associated “Elephant Ears” Those Arising from Healed Fissures or Ulcers or Hemorrhoids with fissures, fistulas, ulcers, hemorrhoids, and non-CD tags. AST of CD can be broad or narrow, single or multiple, small Flat or round Edematous or large, and usually have a smooth surface. These type 1 Soft/compressible Usually hard AST are commonly referred to as “elephant ears.” They are Flesh-colored ϩ/Ϫ waxy Red, blue, or cyanotic generally flesh-colored and painless, may have a cyanotic surface appearance, may or may not be painful, and may have asso- Single or multiple Single or multiple ciated bleeding, are not associated with rectal bleeding, and Varying sizes Usually larger in size than type 1 are mostly associated with CD. The other type of AST, called Painless Often tender or painful type 2, are typically the result of a healed anal fissure, ulcer, Smooth More likely to have irregular surface or hemorrhoid which may also be associated with CD. The No suggestion of Often associated with hemorrhoids presence of ASTs was recorded for each patient and their hemorrhoids features noted from the photographs taken prospectively. The final decision was based on a review of the photographs by a

FIGURE 1. (A) Typical anal skin tags (“elephant ears”). (B) Typical anal skin tags (type 2).

1237 Bonheur et al Inflamm Bowel Dis ● Volume 14, Number 9, September 2008

TABLE 2. Patients with Anal Skin Tags (“Elephant Ears”): abscesses, rectal strictures, and fistulas. Yet AST, acknowl- Ulcerative Colitis Versus Crohn’s Disease (nϭ65) edged to be the most common cutaneous manifestation of CD, have never been studied independently and have infre- Anal Skin Tags quently been even mentioned.4–12 This study is the first to focus specifically on AST in patients with IBD. Crohn’s disease 49 (75.4%) While there have been reports that ASTs are more Ulcerative colitis 16 (24.6%) prominent when CD is active, others have noted that their appearance is independent of the activity of intestinal disease or other perianal manifestations.13,14 This will be the focus of senior IBD endoscopist, an independent senior colon and a future study. rectal surgeon, and the senior and junior authors. Patients AST rarely require treatment and surgical intervention with AST were analyzed for gender, underlying IBD diagno- is generally not recommended. However, Taylor et al15 in- sis, and location in the bowel of the inflammatory process. vestigated excisional biopsies of AST as an adjunct to rectal One patient with indeterminate colitis was excluded from the biopsy in patients with known CD. Granulomas were found study. Data were examined using Fisher’s exact test. A P in almost 30% and, when present, were more plentiful in skin Ͻ 0.05 was a priori considered statistically significant. tags as compared to the biopsies of the rectal mucosa.16 No granulomas were present in control AST obtained from pa- RESULTS tients without CD. This strongly suggests that AST may The most common features of the 2 types of AST are shown provide external evidence of internal disease. Currently, most in Table 1. “Elephant ears” were consistently described as colorectal surgeons avoid excision of AST, fearing compli- being painless, whereas those arising from healed anal fis- cations with wound healing and possible anal inconti- sures or ulcers were more likely to have been associated with nence.1,13 This attitude is likely based on much earlier obser- pain. Examples of typical “elephant ears” are shown in Figure vations which referred to fistulas and abscesses when AST 1A and other less typical AST in Figure 1B. were simply included with other perirectal complications of There were no gender differences within our population CD and not studied independently.6,7 of 169 patients: 56.5% of patients were male and 43.5% were With respect to location of the primary bowel disease female. As determined by accepted clinical, radiological, and the presence of perianal disease in CD, patients with CD endoscopic, and histologic criteria, independent of the AST, limited to the colon have been reported to be more likely to 61.8% of the patients had CD and 37.6% had UC. In all, 65 develop these lesions and to have them as presenting symp- out of 169 (38.5%) patients were determined to have typical toms as compared to patients with small bowel disease (52% ASTs of CD. Of these patients with AST, 75.4% had CD and versus 14%).12 Some have reported an even higher incidence 24.6% had UC (P ϭ 0.005) (Table 2). Location of disease of rectal involvement when perianal disease was present.13 was also reviewed in a subset of patients with AST for which In a patient with a clinical presentation or family his- this information was clearly available (n ϭ 49). Categories tory that may be suggestive of IBD, the presence of AST used were 1) ileitis, 2) colitis, and 3) ileocolitis. A trend was must remain suspect of underlying CD. The question remains noted toward a lower incidence of AST in patients with ileitis whether the diagnosis of CD versus UC should depend more (36.7%) and ileocolitis (16.3%) as compared with 46.9% in on the presence of characteristic AST or the endoscopic patients with colitis (P ϭ 0.067) (Table 3). appearance when that favors the diagnosis of UC. In our study, typical AST were found more frequently DISCUSSION in patients with CD as compared to patients with UC, con- In most cases, intestinal and extraintestinal manifestations of firming previous reports that they are more diagnostic of the CD are identified coincidentally. However, cutaneous signs former. In addition, AST were more commonly found in often precede intestinal disease by months or even years, patients with CD limited to the colon as compared to patients underscoring their diagnostic importance.3–5 Numerous stud- with ileitis or ileocolitis. More details remain to fully char- ies of anorectal manifestations of CD have included fissures, acterize AST of CD and perhaps aid in understanding the prognosis for these affected patients. What are the range of sizes and the variety of appearances? Are they histologically TABLE 3. Distribution of Intestinal Involvement in Subset of different than those which might accompany UC? Attempts ϭ Patients with Type 1 Anal Skin Tags (“Elephant Ears”) (n 49) have already been made to determine the association of Colitis 23 (46.9%) perirectal CD (fistulas and strictures) and the location of the Ileitis 18 (36.7%) inflammatory process in the bowel as well as internal fistu- Ileocolitis 8 (16.3%) las.17,18 What happens to these skin tags coincident with clinical response to specific therapies?

1238 Inflamm Bowel Dis ● Volume 14, Number 9, September 2008 Anal Skin Tags in IBD

As it is not the standard of care to biopsy or remove 3. Morales MA, Marini M, Caminero M, et al. Perianal Crohn’s disease. Int AST in order to make a histologic diagnosis, particularly in J Dermatol. 2000;39:616–623. 4. Keighley MRB, Allan RN. Current status and influence of operation on asymptomatic patients, we must defer to clinical expertise to perianal and colonic disease. Int J Colorect Dis. 1986;1:104–107. make the diagnosis. This study emphasizes the importance of 5. Homan WP, Tang CK, Thorbjarnarson B. Anal lesions complicating carefully describing the morphologic characteristics of AST Crohn’s disease. Arch Surg. 1976;111:1333–1335. 6. Manheim SD. Anorectal complications in regional ileitis. J Mt Sinai of CD to help the clinician make the distinction on physical Hosp N Y. 1955;22:184–186. examination. Perhaps it may be proven that the current ther- 7. Manheim SD, Alexander RM. Anorectal diseases—diagnosis and office apy of CD might once again permit biopsies of AST to be management. J Gastroenterol. 1957;28:403. taken without provocation of disease activity and this will 8. Nivatvongs S. Crohn’s disease. In: Gordon PH, Nivatvongs S, eds. Colon, and Anus. St. Louis, MO: Quality Medical Publishing; assist in establishing a more definitive mechanism for their 1992. p 733. origin. 9. Keighley MR, Williams NS. Treatment of perianal Crohn’s disease. In: ASTs are often encountered on rectal exams, although Surgery of the Anus, Rectum and Colon, vol. 2. Philadelphia: WB Saunders; 1993. p 1807–1809. their diagnostic significance may not even be considered. We 10. Lockhart-Mummery HE. Crohn’s disease: anal lesions. Dis Colon Rec- believe that this has been an important omission in reports on tum. 1975;18:200–202. CD over the years and we hope that this study will serve as 11. Aronoff JS, Korelitz BI, Sohn N, et al. Anorectal Crohn’s disease. a springboard for future investigation of this subject. Indeed, BioDrugs. 2000;13:95–105. 12. Williams DR, Coller JA, Corman ML, et al. Anal complications in AST might represent a genetic phenotype for more accurate Crohn’s disease. Dis Colon Rectum. 1981;24:22–24. classification of CD. For now, we encourage clinicians to 13. Triantafillidis JK, Emmanouilidis A, Nicolakis D, et al. Perianal Crohn’s review the descriptive characteristics of typical AST of CD disease. Clinical features and follow-up of 33 Greek patients. Hellenic J Gastroenterol. 1997;10. presented in this article and when they are encountered on 14. Funayama Y, Sasaki I, Imamura M, et al. Surgical management of physical exam to consider the diagnosis of CD even before perianal lesions in Crohn’s disease. Nippon Geka Gakkai Zasshi. 1987; any endoscopic or radiologic investigation is pursued. This is 88:562–568. 15. Taylor BA, Williams GT, Hughes LE, et al. The histology of anal skin particularly so in young people with diarrhea, abdominal tags in Crohn’s disease: an aid to confirmation of the diagnosis. Int J pain, retarded growth, and development and/or a family his- Colorectal Dis. 1989;4:197–199. tory of CD. 16. Korelitz BI, Sommers SC. Rectal biopsy in patients with Crohn’s disese. JAMA. 1977;237:2742–2744. REFERENCES 17. Sachar DB, Bodian Ca, Goldstein ES, et al. Task Force on Clinical Phenotyping of the IOIBD. Is perianal Crohn’s cisease associated with 1. AGA technical review on perianal Crohn’s disease. Gastroenterology. intestinal fistulization? Am J Gastroenterol. 2005;100:1547–1549. 2003;125:1508–1530. 18. Fields S, Rosainz L, Korelitz BI, et al. Rectal strictures in Crohn’s 2. Sohn N, Weinstein M, Robbins R. Anorectal disorders. Curr Probl Surg. disease and co-existent perirectal complications. Inflamm Bowel Dis. 1983;20:1–66. 2008;14:29–31.

1239